This article presents a robust and computationally inexpensive technique of component-level fault detection in aircraft gas-turbine engines. The underlying algorithm is based on a recently developed statistical pattern recognition tool, symbolic dynamic filtering (SDF), that is built upon symbolization of sensor time series data. Fault detection involves abstraction of a language-theoretic description from a general dynamical system structure, using state space embedding of output data streams and discretization of the resultant pseudo-state and input spaces. System identification is achieved through grammatical inference based on the generated symbol sequences. The deviation of the plant output from the nominal estimated language yields a metric for fault detection. The algorithm is validated for both single- and multiple-component faults on a simulation test-bed that is built upon the NASA C-MAPSS model of a generic commercial aircraft engine.
BroersenP.Estimation of parameters of non-linear dynamical systems. Int. J. Non-linear Mech., 1974, 9, 355–361.
2.
Van LithP.WitteveenH.BetlemB.RoffelB.Multiple nonlinear parameter estimation using pi feedback control. Control Eng. Pract., 2001, 9, 517–531.
3.
WanE.Van der MerweR.The unscented kalman filter for nonlinear estimation. In Proceedings of the IEEE Symposium 2000 ASSPCC, Lake Louise, Alta, Canada, 1–4 October 2000, 153–158.
4.
JulierS.UhlmannJ.A new extension of the kalman filter to nonlinear systems. Proc. SPIE, 1997, 3068, 182–193.
5.
ChingJ.BeckJ.PorterChinK.Bayesian state and parameter estimation of uncertain dynamical systems. Prob. Eng. Mech., 2006, 21(1), 81–96.
6.
BremerC.KaplanD.Markov chain Monte Carlo estimation of nonlinear dynamics from time series. Physica D, 2001, 160, 116–126.
7.
WangY.GengL.Bayesian network based fault section estimation in power systems. 14–17 November 2006, IEEE Region 10 Annual International Conference, TENCON. Hong Kong, China.
8.
HoffmanA.van der MerweN.The application of neural networks to vibrational diagnostics for multiple fault conditions. Comput. Stand. Interfaces, 2002, 24(2), 139–149.
9.
DavidB.BastinG.Parameter estimation in nonlinear systems with auto and crosscorrelated noise. Automatica, 2002, 38, 81–90.
10.
GhanemR.RomeoF.A wavelet-based approach for model and parameter identification of non-linear systems. Int. J. Non-linear Mech., 2001, 36, 835–859.
11.
YaoL.SetharesW.Nonlinear parameter estimation via the genetic algorithm. IEEE Trans. Signal Process., 1994, 42(4), 927–935.
12.
RayA.Symbolic dynamic analysis of complex systems for anomaly detection. Signal Process., 2004, 84(7), 1115–1130.
13.
RajagopalanV.RayA.Symbolic time series analysis via waveletbased partitioning. Signal Process., 2006, 86(11), 3309–3320.
14.
GuptaS.RayA.SarkarS.YasarM.Fault detection and isolation in aircraft gas turbine engines. Part 1: Underlying concept. Proc. IMechE, Part G: J. Aerospace Engineering, 2008, 222(3), 307–318.
15.
SarkarS.YasarM.GuptaS.RayA.MukherjeeK.Fault detection and isolation in aircraft gas turbine engines. Part 2: Validation on a simulationtest bed. Proc. IMechE, Part G: J. Aerospace Engineering, 2008, 222(3), 319–330.
16.
SarkarS.RaoC.RayA.Statistical estimation of multiple faults in aircraft gas turbine engines. Proc I MechE Part G: J. Aerospace Engineering, 2009, 223(4), 415–424.
17.
RaoC.RayA.SarkarS.YasarM.Review and comparative evaluation of symbolic dynamic filtering for detection of anomaly patterns. 2008, DOI 10.1007/s11760-008-0061-8.
18.
SubbuA.RayA.Space partitioning via Hilbert transform for symbolic time series analysis. Appl. Phys. Lett., 2008, 92(08), 084107–084107.
19.
FrederickD.DeCastroJ.LittJ.Users guide for the commercial modular aero-propulsion system simulation (c-mapss). October 2007, NASA/TM2007-215026.
20.
C-MAPSS: Commercial modular aero-propulsion system simulation, NASA GRC Software Repository, 2007.
21.
KobayashiT.SimonD.A hybrid neural network-genetic algorithm technique for aircraft engine performance diagnostics. 8–11 July 2001, Proceedings of the 37th Joint Propulsion Conference and Exhibit co-sponsored by the AIAA, ASME, SAE, and ASEE. Salt Lake City, Utah.
22.
HopcroftJ.MotwaniR.UllmanJ.Introduction to automata theory, languages, and computation, 3rd edition. Boston, Massachusetts, USA, Addison-Wesley Longman Publishing Co., Inc, 2006.
23.
KokarM.On consistent symbolic representations of general dynamic systems. IEEE Trans. Syst. Man Cybernetics, 1995, 25(8), 1231–1242.
24.
LevyL.Discrete structures of computer science. New York, John Wiley & Sons, 1980.
GautamaT.MandicD.HulleM. V.A differential entropy based method for determining the optimal embedding parameters of a signal. In Proceedings of the. IEEE International Conference on Acoustics, speech, and signal processing, Hong Kong, China, Vol. 6 (6–10) pp. VI-29-32, 6–10 April 2003.
27.
BeirlantJ.DudewiczE. J.GyrfiL.MeulenE. C.Nonparametric entropy estimation: An overview. Int. J. Math. Stat. Sci., 1997, 6, 17–39.
28.
MartinsJ.DenteJ.PiresA.MendesR.Language identification of controlled systems: Modeling, control, and anomaly detection. IEEE Trans. Syst. Man Cybernetics, Part C: Applications And Reviews, 2007, May.
29.
NarendraK.ThathacharM.Learning automata an introduction. Upper Saddle River, New Jersey, USA, Prentice-Hall, 1989.
30.
KobayashiT.SimonD. L.Hybrid kalman filter approach for aircraft engine in-flight diagnostics: Sensor fault detection case. J. Eng. Gas Turbines Power, 2007, 129, 746–754.
31.
DudaR.HartP.StorkD.Pattern classification. New York, John Wiley & Sons Inc, 2001.
32.
SimonD. L.GargS.Optimal tuner selection for kalman filterbased aircraft engine performance estimation. J. Eng. Gas Turbines Power, 2010, 132(3), 031601–031601.
33.
LuP.ZhangM.HsuT.ZhangJ.An evaluation of engine faults diagnostics using artificial neural networks. J. Eng. Gas Turbines Power, 2001, 123(2), 340–346.
34.
PoorV.An introduction to signal detection and estimation, 2nd edition. New York, USA, Springer-Verlag, 1988.
35.
YangT.Computational verb theory: Ten years later. Int. J. Comput. Cognit., 2007, 5, 63–86.
